Jump to navigation Jump to search
Planning a Co-operative Cloud
Can we build a co-operative cloud service, what would it look like? Join up in #coopcloud on Slack.
- Very-PC — Sheffield based, try to be as green as possible, not much Linux experience but very supportive
- Broadberry — London based
- Avensys] — Shefield based, very helpful, money for electricity goes to Good Energy, 24 hour access, cycling distance for Webarchitects techies to swap out failed disks etc
Notes from a discusion at Wortley Hall in 2016
- A real space for Cooperatives to run their own infrastructure.
- Briefly considered what the USP for cooperatives running cloud stuff.
- As well as the ethics of "keeping it in the family" what are the technical advantages?
- What would be the sell to the rest of the tech community at large?
- Seems a minimum would be:
- Some kind of VM service
- Some kind of relational database service
- Some kind of S3 compatible large scale storage
- There is certainly software in this category of "create your own cloud" - for example OpenStack.
- Looking into this seems a wise start.
- Should prioritise "code as infrastructure" and encourage people to use tools like Salt/Ansible/Terraform to provision infrastructure which they keep under close version control.
- Ride the wave of popularity around this stuff in the industry
- Avoid development costs of creating a complex user interface for the code from the offset though this might have to come later.
- Obviously Outlandish have a deal of experience around this UI stuff but feels this should be a "headless" offer to devops people first.
- Bring a "developer first" tool to allow uptake
- Drop in plugins for Ansible/Terraform or AWS compatiable APIs that would enable existing infrastructure as code stuff to be dropped in with supreme developer ease into existing projects to drive developer adoption.
- Chris says Web Architects have a simple infrastructure as code offer which is simple keeping DNS records in GitLab
- Alex said Outlandish use Ansible for provisioning pretty extensively because it was a good gateway for people who hadn't used anything before into this space
- Noted he'd love to just change a few variables around and have the infrastructure on CoopCloud rather than AWS
- We should try for an MVP of the VM/DB/S3 offer, try it internally and see what the commercial viability is.
- Hardware should obviously be looked at, consider commodity based simple chassis (a la the "big boys") versus expensive specialist hardware like blade servers etc.
- Because of hardware limitations the service should transparently build capacity off other commercial clouds so the user never experiences this capacity problem.
- There is probably an opportunity to get in on the containerisation stuff while it is in its infancy and make cooperatives a leader in this space in the UK.